---
dataset_info:
  features:
  - name: id
    dtype: int64
  - name: multimodal_question
    dtype: string
  - name: answer
    dtype: string
  - name: rationale
    dtype: string
  - name: text_only_question
    dtype: string
  - name: image_source
    dtype: string
  - name: evidence
    dtype: string
  - name: resolution
    dtype: string
  - name: proportion_of_roi
    dtype: string
  - name: category
    dtype: string
  - name: text_in_image
    dtype: string
  - name: rationale_granularity
    dtype: string
  - name: image
    dtype: image
  - name: cropped_image
    dtype: image
  splits:
  - name: train
    num_bytes: 515536503.0
    num_examples: 500
  download_size: 515227519
  dataset_size: 515536503.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---